Stone, T.W., Lui, C. and Addae, J.I. (2010) Effects of ethylenediamine – a putative GABA-releasing agent – on rat hippocampal slices and neocortical activity in vivo. European Journal of Pharmacology, 650 (2-3). pp. 568-578. ISSN 0014-2999. http://eprints.gla.ac.uk/46340/ Deposited on: 22 November 2010

Enlighten – Research publications by members of the University of Glasgow http://eprints.gla.ac.uk

Effects of ethylenediamine – a putative GABA-releasing agent – on

rat hippocampal slices and neocortical activity in vivo

Trevor W Stone1 , Caleb Lui1 and Jonas I. Addae2

1 Neuroscience and Molecular Pharmacology,

Faculty of Biomedical and Life Sciences, University of Glasgow,

Glasgow G12 8QQ, U.K.

2 Dept of Preclinical Sciences,

St. Augustine Campus

University of the West Indies,

Trinidad and Tobago.

Correspondence: Prof T. W. Stone, West Medical Building, University of Glasgow,

Glasgow G12 8QQ, UK.

Tel - +44 141 330 4481; fax - +44 141 330 5481;

e-mail - Trevor.Stone@glasgow.ac.uk

1

ABSTRACT

The simple diamine diaminoethane (ethylenediamine, EDA) has been shown to

activate GABA receptors in the central and peripheral nervous systems, partly by a

direct action and partly by releasing endogenous GABA. These effects have been

shown to be produced by the complexation of EDA with bicarbonate to form a

carbamate. The present work has compared EDA, GABA and β-alanine responses in

rat CA1 neurons using extracellular and intracellular recordings, as well as

neocortical evoked potentials in vivo. Superfusion of GABA onto hippocampal slices

produced depolarisation and a decrease of field epsps, both effects fading rapidly,

but showing sensitivity to blockade by bicuculline. EDA produced an initial

hyperpolarisation and increase of extracellular field epsp size with no fade and only

partial sensitivity to bicuculline, with subsequent depolarisation, while β-alanine

produces a much larger underlying hyperpolarisation and increase in fepsps,

followed by depolarisation and inhibition of fepsps. The responses to β-alanine, but

not GABA or EDA, were blocked by strychnine. In vivo experiments, recording

somatosensory evoked potentials, confirmed that EDA produced an initial increase

followed by depression, and that this effect was not fully blocked by bicuculline.

Overall the results indicate that EDA has actions in addition to the activation of GABA

receptors. These actions are not attributable to activation of β-alanine-sensitive

glycine receptors, but may involve the activation of sites sensitive to adipic acid,

which is structurally equivalent to the dicarbamate of EDA. The results emphasise

the complex pharmacology of simple amines in bicarbonate-containing solutions.

Key-words: Ethylenediamine; 1,2-diaminoethane; GABA; bicuculline; kynurenic

acid; hippocampus; cortical evoked potentials

2

1. INTRODUCTION

Ethylenediamine (1,2-diaminoethane, EDA) is used as a complexing and

solubilising agent in preparations such aminophylline, which is a complex of EDA and

theophylline. When aminophylline was applied by microiontophoresis to

spontaneously firing neurones in vivo (Stone and Perkins, 1979) it produced a

depression of neuronal firing, leading to the realisation that the inhibition was entirely

attributable to the EDA, not the theophylline component of the complex (Perkins and

Stone, 1980,1982a; Perkins et al., 1981). This conclusion was confirmed by others

studying neuronal activity in the hippocampus (Blaxter and Cottrell 1985), cerebellum

and invertebrate CNS (Bokisch et al., 1984).

Both EDA and GABA showed the same reversal potential, the depressant activity

was blocked by bicuculline (Perkins et al., 1981; Perkins and Stone, 1982a) and

potentiated by benzodiazepines (Bokisch et al., 1984). Neurochemical analyses

showed that EDA could be transported into GABA-containing synaptic terminals,

provoking the selective release of GABA but not dopamine (Lloyd et al., 1982a,b;

Kerr and Ong, 1982, 1987; Sarthy, 1983). Overall, EDA was a GABA-mimetic, acting

partly indirectly by GABA release (either by uptake and displacement of synaptic

GABA or by promoting the exchange of GABA for EDA) and partly by direct

activation of GABAA receptors (Morgan and Stone, 1982; Davies et al., 1982). Later

work indicated that EDA can activate GABA receptors in smooth muscle (Ong and

Kerr, 1987; Barbier et al, 1989; Maggi et al., 1989; Chaudhuri and Ganguly, 1994,

Begg et al., 2002), partly via endogenous GABA release (Kerr and Ong, 1984; Erdo

et al., 1986; Mckay and Krantis, 1991).

In order to explain this neuronal inhibition by a simple amine lacking any oxygen-

containing functional group, it was proposed that the positively charged amine groups

3

of EDA might interact with bicarbonate ions in solution (Stone and Perkins, 1984) to

generate an aliphatic structure resembling GABA. The existence of the postulated

complex was subsequently confirmed and identified as EDA monocarbamate (Curtis

and Malik, 1984; Kerr and Ong, 1987), which reproduced the activity of EDA.

Analogues with longer or branched structures are much less effective as GABA-

mimetics (Perkins and Stone 1982a).

The similarity between EDA and GABA pharmacology extended to the ability of

nipecotic acid to inhibit EDA uptake in brain slices (Lloyd et al., 1982b) and the ability

of depolarising stimuli to release EDA in a calcium-dependent fashion.

Overall, these results indicated that EDA might make a valuable tool to explore

aspects of GABA synaptic pharmacology, promoting the effects of endogenous

GABA released from synaptic terminals. However, there remain several unanswered

questions. For example, GABA acts on CA1 apical dendrites with a hyperpolarisation

that is not blocked by bicuculline (Alger and Nicoll, 1982) although it is still mediated

via GABAA and not GABAB receptors. Therefore, the present study was designed to

re-examine the pharmacology of EDA under the controllable conditions of

hippocampal slices. In addition, it was of interest to examine the effects of EDA in

vivo, where the intercellular milieu is more complex and adaptive than under in vitro

conditions.

2. MATERIALS AND METHODS

2.1 General

All procedures performed in this work were in accordance with the regulations and

recommendations of the Animals (Scientific Procedures) Act, 1986 of the United

4

Kingdom Home Office. Hippocampal slices were prepared as described previously

(see Stone, 2007; Ferguson and Stone, 2008). Briefly, male Wistar rats weighing

100-150g were killed by administering an overdose of urethane (10ml/kg body weight

i.p., as a 25% solution in water) followed by cervical dislocation. The brain was

rapidly removed into ice-cold artificial cerebrospinal fluid (aCSF) of composition: (in

mM) NaCl 115; KH2PO4 2.2; KCl 2; MgSO4 1.2; NaHCO3 25; CaCl2 2.5; D-

glucose 10, gassed with 5%CO2 in oxygen. The hippocampi were chopped into

450µm transverse slices using a McIlwain tissue chopper. The slices were

preincubated at room temperature for at least 1 hour in a water-saturated

atmosphere of 5%CO2 in O2 before individual slices were transferred to a 1 ml

capacity superfusion chamber for recording.

For extracellular recording, submerged slices were superfused with aCSF at 28-

30°C and a flow rate of 3-4 ml/min. The Schaffer collaterals and commissural

afferents to the CA1 region were stimulated using a concentric bipolar electrode

(Harvard Apparatus, Edenbridge, UK) positioned in the stratum radiatum, (0.1 Hz

stimulation using a pulse duration of 50-300µs. The stimulating electrode tip was

located immediately internal to the stratum pyramidale at the border between the

CA1 and CA2 regions. For antidromically-evoked potentials, the stimulating electrode

was placed in contact with the alvear white matter containing the axons of the

pyramidal neurons. Recording electrodes were constructed from fibre-containing

borosilicate glass capillary tubing (Harvard Apparatus, Edenbridge, Kent, UK), with

the tips broken back under microscopic control to 2-4µM, DC resistance

approximately 5MΩ. These electrodes were filled with a solution of 1M NaCl. The

recording electrode tip was placed either in the stratum pyramidale (for population

spike recordings) or internal to the CA1 pyramidal cell layer at the point of maximum

5

epsp amplitude. The potentials were amplified and captured on a micro1401 interface

(CED, Cambridge Electronic Design, Cambridge, UK) for storage on computer and

subsequent analysis using Signal software (CED, Cambridge, UK). The amplitude of

population spikes was taken as the potential difference between the negative and

positive peaks of the potential. The negative slope of the epsp was taken as the

gradient of the line of best fit of all sampled points between approx. 25% and 75%

after the start of the negative slope. The epsp amplitude was taken as the potential

difference between the negative peak of the epsp and the pre-stimulus baseline.

Intracellular recordings were made as described previously (Stone, 2007), using

sharp electrodes pulled from fibre-containing glass capillary tubing (Clark

Electromedical, Reading, UK, and Harvard Apparatus, Edenbridge, UK) using a

vertical Narashige puller. The electrodes were filled with 1M potassium acetate, and

had tip resistances of 90-120 MΩ. Potentials were amplified via an Axoclamp-2

system or Neurolog DC amplifier in bridge balance mode, with a filter bandwidth

between DC and 500Hz to reduce higher frequency components of epsp recordings.

Following penetration, cells were allowed to reach a maximum resting potential and

then left for 30 min before the addition of compounds into the superfusion line. Cells

were only used if they exhibited stable resting potentials of at least –60mV. Neuronal

excitability was tested by applying intracellular depolarising current pulses between

0.05 and 1.0nA amplitude. Input resistance was monitored using 0.05 to 1.0nA

hyperpolarizing pulses. Responses were digitised at 20kHz via a CED (Cambridge

Electronic Design, Cambridge, UK) micro1401 interface and stored on computer for

later analysis using the Signal programme (CED). Only one cell was studied in each

slice, to preclude the possibility that drug superfusion might alter the responsiveness

of subsequent neurons in the same slice.

6

2.2 In vivo electrophysiology

The in vivo work was approved by the local Ethics Committee for the use of

animals. Male Sprague Dawley rats weighing 200-320g were housed and maintained

at the animal care facility of the university’s veterinary hospital according to the

standards of the local ethics committee. The rats had free access to food and water

with a 12 hour light/dark cycle. The experimental procedures we used for recording

somatosensory evoked potentials (SEPs) have been described in previous reports

(Addae et al., 2000, 2007). Briefly, following anesthesia with urethane (1.5-1.7 g/kg)

the rats were mounted in a stereotaxic frame and the body temperature maintained

at 37oC by means of a heating blanket and a rectal thermistor. A burr hole, 3-4 mm in

diameter, was made to expose the cortical area that represents the forepaw and the

dura mater and arachnoid mater were removed in this area to allow recording of the

SEPs and application of drugs via a cortical cup. The contralateral forepaw was

stimulated at 0.5 Hz using needle electrodes inserted into the skin of the forepaw,

and the SEPs were recorded at 0.8-1.2mm below the cortical surface with a glass

microelectrode filled with 3M NaCl (resistance of approximately 5 MΩ). Individual

SEPs were recorded using a CEDmicro1401 computer interface and Signal software

(Cambridge Electronic Design, UK). Changes in SEP amplitude were calculated as a

percentage of the mean baseline value. Following application of drugs the cortex was

washed by continuous superfusion of pre-gassed (95%O2/ 5%CO2 ) artificial

cerebrospinal fluid (aCSF) with the following composition (in mM): NaCl 115.0; KCl

2.0; KH2PO4 2.2; NaHCO3 25.0; D-Glucose 10; MgSO4 1.2 and CaCl2 2.5. 2.5. EDA

and bicuculline methobromide were dissolved in the aCSF, although when EDA was

7

being applied at high concentrations it was dissolved in aCSF with reduced

concentrations of NaCl to maintain isotonicity of the solution.

Data are expressed as mean ± S.E.M.. Statistical comparisons of two data sets

were made with an unpaired Student’s t-test. Comparisons of multiple data sets were

made using ANOVA.

2.3 Statistics

Data are presented throughout as mean ± standard error. Baseline values were

obtained from a stable 10min period of evoked potential size prior to the addition of

any drugs, with the first of those potentials being defined as 100%. This allowed the

10min pre-drug period to provide an indication of baseline variance. For statistical

comparisons between two groups, unpaired, two-tailed t-tests were used, but where

two or more data sets were compared with a common control, an analysis of

variance (ANOVA) was performed, followed by Dunnett’s test for comparisons with a

common control, or the Bonferroni test for selected pairs of columns as appropriate.

All analyses were conducted using Instat software.

2.4 Sources

Glutamate, 4-aminobutyric acid (GABA), glycine, β-alanine, adipic acid, α-amino-

adipic acid, ethylenediamine dihydrochloride (EDA), strychnine and nipecotic acid,

were obtained from Sigma-Aldrich Chemical Company, Poole, UK.

(-)bicuculline methobromide (BIC), 3-((R)-2-carboxypiperazin-4-yl)-propyl-1-

phosphonic acid (CPP), (3-aminopropyl)(diethoxymethyl)phosphinic acid

(CGP35348) and gabazine were obtained from Tocris Chemicals, Bristol, UK.

8

3. RESULTS

3.1 Responses to EDA and GABA

The main response to EDA was a reduction in the size of fepsps (Fig. 1A,B). The

time course of responses is illustrated by the family of curves obtained in response to

maintained (10min) applications of EDA such that responses reached a plateau

(Fig.1A). The corresponding log concentration-response curves are constructed in

Fig.1C. When perfusion with EDA was maintained until this plateau of depression, a

significant effect could be seen at a concentration as low as 50µM (Fig. 1B,C). None

of these responses to EDA were associated with any changes in the size of the

axonal volley (Fig.1B). In an initial series of 16 slices tested with EDA at 1mM or

above, 4 slices showed a depression of fepsps which was preceded by a small

increase in size which reached 120.9 ± 2.2% of the initial baseline (n = 4, significantly

different from baseline, P = 0.006), and was not clearly related to concentration.

At concentrations below approximately 1mM, the effects of EDA were normally

fully reversible, with potential size returning to levels not significantly different from

the initial baseline values (Fig. 1A). At concentrations of 1 mM or above, a proportion

of slices showed a long-lasting depression of fepsp size. Overall, of the initial group

of 16 slices perfused with EDA alone at 1mM or above, 12 showed a fully reversible

depressant effect of EDA on fepsps, while 4 showed persistent depression after

recovering only partially towards baseline levels. This long-lasting depression was

even more evident when slices were perfused for a shorter period of 5min (Fig. 1D).

For comparison, GABA was superfused over slices at a range of concentrations

similar to those used for EDA (Fig. 1C). The responses obtained were of a broadly

similar magnitude and time course to those seen with EDA, although there were no

9

instances of initial increases in fepsp size, and all responses returned to the baseline

levels, with no maintained depression. Indeed, when applied at concentrations of

500µM or above for 10min, the fepsp size reached a peak of depression after about

1min, before this effect began to fade, starting to recover towards baseline values

during the continued presence of GABA. In these cases, for the purpose of

constructing the concentration-response curve in Fig. 1C, the amplitude of the

response was taken as the maximum extent of the depression.

3.2 GABA receptor blockade

In order to examine the pharmacology of these responses, EDA was applied at a

concentration of 1mM for periods of 2 min, once every 10-15 min. This protocol

allowed the establishment of two consistent control responses, followed by the

application of a potential modulator and then recovery of the original EDA response.

The results were discarded from any slices in which the EDA response did not

recover to within 10% of its initial value.

The application of bicuculline methobromide to slices tends to result in the

development of multiple population spikes in the recordings of fepsp, together with

the occurrence of spontaneous epileptiform bursting. These experiments were

therefore conducted in slices in which a cut was made between the CA3 and CA1

regions, approximately at the border between CA2 and CA3. While this did not

completely prevent signs of hyperexcitability, the changes seen did not interfere with

the overall behaviour of the slices, or affect their eventual recovery.

Slices were superfused with bicuculline for 10 min before the addition of 1mM

EDA for 2 min, with continued perfusion of bicuculline during subsequent applications

of EDA. The application of bicuculline at a concentration of 20µM produced a slowly

10

developing reduction in the size of fepsp responses to EDA, with a maximal reduction

after approximately 30 min of bicuculline (Fig. 2A), and with no changes of axonal

volley amplitude (Fig. 2B). A higher concentration of 100µM bicuculline antagonised

EDA more rapidly but did not produce any greater maximal degree of antagonism

(Fig. 2C).

Since bicuculline is known to affect some receptors and channels in addition to

those for GABA, experiments were also performed using gabazine, a highly selective

antagonist at GABAA receptors. In these cases, a similar, partial antagonism of EDA

was seen when gabazine (20µM) was used, a concentration known to block GABAA

receptors (Lindquist et al., 2005) (Fig. 2C). A higher concentration of 100µM did not

produce a significantly greater reduction of the EDA response (Fig. 2C).

In contrast to these results, bicuculline at 20µM produced a complete blockade of

the responses to GABA (1mM; data not shown). On the other hand, neither

strychnine (10 or 100µM) (Fig. 2C) nor the GABAB receptor antagonist CGP35348

(100µM; data not shown) were able to produce any significant reduction in the size of

responses to EDA.

Similarly, perfusion with neither bicuculline nor strychnine produced any change in

the size of the initial increase of fepsp size. On those slices showing this aspect of

the response, the fepsp size was increased by 16.43 ± 1.20% with EDA alone (n =

3); 16.87 ± 2.89% with EDA and bicuculline (n = 3); and 16.77 ± 2.5% with EDA and

strychnine (n = 3). These increases were not significantly different (ANOVA (F(2,8) =

0.0096; P = 0.99).

3.3 Nipecotic acid

11

The GABA transport (GAT-1) inhibitor nipecotic acid (100µM) was applied for

10min before and during a series of applications of EDA. The concentration of 100µM

was selected as one which had no effect alone on the fepsp size. With this protocol,

nipecotic acid produced a gradual enhancement of the size (Fig. 3A,B,C) and

particularly the duration of responses to EDA, with no change in the size of the

axonal volley (Fig. 3B). The potentiation of EDA responses recovered to a level not

significantly different from the original values (Fig. 3A,C,D)

3.4 Chelation

EDA is well-established as a chelator of metal ions, for which property it is widely

used in the chemical synthesis of organo-metallic complexes, it was considered that

the depression of evoked potentials, particularly that component of the depression

that was not prevented by bicuculline or gabazine, might reflect the chelation of Ca++

ions. Doubling the calcium content of the aCSF increased the size of the baseline

fepsp by approximately 20%, but did not reduce the percentage fepsp depression

produced by 500µM EDA (41.6 ± 3.7%, n = 4 in normal aCSF, 38.8 ± 1.7%, n = 4 in

elevated calcium, P = 0.53) . Similarly the possibility was also considered that the

initial increase in fepsp slope seen in some slices (see above) might result from the

chelation of Mg++, possibly resulting in an initial overdepolarisation of slices caused

by increased sensitivity of NMDA receptors. When a slice was encountered in which

EDA produced a clear and reproducible initial increase of fepsp size, the selective

NMDA receptor antagonist CPP (10µM) was superfused. However, CPP did not

prevent either the early hyperexcitability (control increase 18.23 ± 3.6%, with CPP

22.8 ± 3.8%, n = 3, P = 0.44) or the predominant depression of fepsp slope (control

depression 50.9 ± 5.6%, with CPP 44.9 ± 6.3, n = 4, P = 0.51).

12

3.5 Antidromic potentials

The foregoing results suggest thatsome of the depression of fepsps produced by

EDA could be attributed to activation of GABAA receptors, but the origin of the

occasional increase of excitability, and of that fraction of neuronal inhibition that was

not blocked by bicuculline, remained unclear. It was therefore decided to test EDA on

non-synaptic potentials evoked by the antidromic activation of CA1 pyramidal cells.

Even with this paradigm, EDA at 300µM continued to depress the amplitude of the

antidromic population spikes (Fig. 4A,B). When these slices were superfused with

kynurenic acid, a non-selective antagonist at glutamate receptors (Perkins and

Stone, 1982b; Stone 2001), using a concentration of 2mM, there was no significant

change in the size of the EDA-induced depression (Fig. 4A,B).

3.6 β-alanine

In addition to forming a GABA-like agonist complex, the structure of EDA might

lend itself to the formation of a complex or carbamate compound similar in electronic

structure to β-alanine, which has inhibitory activity on central neurons. Applications of

β-alanine depressed orthodromic potentials with a potency comparable to that of

EDA (Fig. 5A), producing a reduction in the fepsp of 83.05 ± 3.48% (n = 5) at a

concentration of 2mM and 44.72 ± 3.94% (n = 4) at 500µM. Superfusion with

bicuculline at 100µM did not change the depression of fepsps whereas strychnine

(20µM) produced a selective block of the depressant response which developed

during a strychnine application lasting 30min (Fig. 5A,B).

The depression of fepsps by β-alanine was preceded in most cases (7 of 8 slices

tested using 500µM) by a significant increase in the slope of the potential (Fig. 5A).

13

This was larger, longer-lasting and appeared in a larger proportion of slices than was

the initial increase sometimes seen with EDA. The initial increase of fepsp size was

not reduced by superfusion with bicuculline or strychnine (Fig. 5A,C).

3.7 Intracellular recordings

Since the foregoing results presented a complicated picture of mixed excitatory

and inhibitory components in the response to EDA, experiments were performed to

address further the mechanism of action of EDA by intracellular recordings using

classical sharp electrodes. A total of 22 neurones in the CA1 pyramidal layer were

recorded (in separate slices), with resting membrane potentials ranging from -66 to -

81mV. The application of EDA to the slices for 2min produced, in most cases (16 of

22), and at both the 300µM or 1mM concentrations examined, a rapid depolarisation

of the neurones (Fig. 6A).

The depolarisations produced by EDA at 1mM reached 27.93 ± 2.0 mV (n = 16).

This depolarisation was associated with a marked decrease of membrane input

resistance from baseline values of 57.9 ± 3.8 MΩ to 32.3 ± 2.5 MΩ (P = 0.0002; n =

6)(Fig. 6A,B), the decreased resistance still being apparent when the membrane

potential was clamped at the original resting potential (Fig. 6C). The application of

concentrations greater than 1mM usually resulted in overdepolarisation of the

neurones.

In several cases (6 of 22 cells), the depolarisation was preceded and followed by

a marked hyperpolarisation (Fig. 6B) which reached 9.75mV ± 2.15 in those 6

neurons, all of which were treated with EDA at 1mM.

The mechanism of the EDA-induced depolarisation was probed by perfusing with

kynurenic acid (2mM) for 10 min before, during and for 10min following EDA

14

(500µM). Kynurenic acid produced only small and non-significant changes in the size

of the overall EDA depolarisations, which were reduced from 22.2 ± 3.6 mV to 21.1 ±

3.9 mV (P = 0.83, n = 4).

Superfusion with bicuculline (20µM) reduced the amplitude of EDA depolarisations

elicited at 1mM (Fig. 7A,B) or 300µM concentrations, an effect which reached a

maximum plateau after about 30 min of superfusion, consistent with the effects seen

using synaptically-evoked potentials (above). Strychnine 20µM did not change the

magnitude of the EDA depolarisations (controls 26.6mV ± 3.6mV, with strychnine

24.3 ± 2.9mV, n = 4, P = 0.64).

Applications of GABA for 2 min (1mM) generated complex responses which in

most cases (5 of 7 cells) involved an early depolarisation which attained a peak

within about 30secs and then declined substantially during the presence of GABA.

These responses closely resembled in profile the changes of extracellularly recorded

fepsps described above. The responses were fully blocked by superfusion with

bicuculline (20µM), but not strychnine (20µM), and were unaffected by the GABAB

receptor antagonist CGP35348 (100µM)(data not shown).

When applied during intracellular recording, β-alanine (1mM) produced a

depolarisation of 22.3mV ± 3.1 (n = 4), preceded in 3 cases by an initial and

subsequent hyperpolarisation (6.4mV ± 0.69, n = 3) very similar to those seen with

EDA (above, Fig. 6B). Strychnine produced a gradual blockade of the depolarising

phase of these responses (to 8.9mV ± 2.7, n = 4, P = 0.002 using paired t test), with

no effect on the hyperpolarisations (6.0mV ± 0.67, n = 3, P = 0.72). Bicuculline

(100µM) had no significant effect on either component of the β-alanine responses.

3.8 Adipic acid

15

Given the chemical structure of EDA, and previous evidence for its ability to form

acidic complexes with bicarbonate ions, the possibility was considered that a similar

di-carboxy adduct might be responsible for the bicuculline-resistant neuronal

depolarisation. Such an adduct should most closely resemble adipic acid in structure,

and this compound was therefore tested on slices.

Adipic acid itself produced little effect at 1mM, with small and inconsistent effects

at 2mM. At a concentration of 10mM it produced changes of the fepsps which were

either increased (22.6 ± 3.0%, n = 4, significantly different from baseline, P = 0.005),

or decreased (21.8 ± 3.3%, n = 5, significantly different from baseline, P = 0.003).

These effects were not only smaller than those produced by EDA (at 1mM), but were

also slower in both onset and recovery (Fig. 8A). Intracellular recordings showed

that, like EDA, the dominant effect of adipic acid on CA1 neurones was one of

depolarisation with increased membrane conductance but these responses were

unaffected by bicuculline (20µM) or strychnine (20µM) (data not shown).

In view of the structural similarity between EDA monocarbamate and adipic acid

and α-amino-adipic acid (αAA), and the fact that the latter compound was one of the

earliest antagonists to be identified at glutamate receptors (Biscoe et al., 1977), the

possibility was considered that EDA carbamates and adipic acid might both act as

weak NMDA receptor agonists, and might be antagonised by αAA. However, this

was not the case, and when superfused at a concentration of 1mM, αAA had no

significant activity against the responses to either EDA or adipic acid (Fig. 8A,B).

3.9 Role of extracellular bicarbonate

In order to pursue the question of whether the activity of EDA was attributable to

complex formation with bicarbonate, 25% of the normal bicarbonate content of the

16

aCSF was replaced by HEPES. This substitution reduced significantly the depression

of evoked potentials by 1mM EDA from 75 ± 5% (n = 6) to 40 ± 5% (n = 6) (P <

0.001), while there was no change in the depression produced by 1mM GABA

(controls 40 ± 5%, with HEPES 40 ± 5%, n = 6, n.s.).

3.10 Channel modulation

4,4’-di-isothiocyano-stilbene-2,2’-disulphonic acid (DIDS), a blocker of chloride /

bicarbonate exchange, is known to block the conductance changes induced by

GABA and hence to reduce its depolarising and hyperpolarising actions (Bonnet and

Bingmann, 1995). When tested against responses to EDA, DIDS (20µM) reduced

substantially the depolarisation and depression of evoked potentials: the effect of

1mM EDA was reduced from 70 ± 5% (n = 8) to 20 ± 5% (n = 8) (P < 0.001),

consistent with a similar site and mechanism of action between EDA and GABA.

3.11 In vivo experiments

In order to determine whether EDA had any effect on evoked potentials under more

physiological conditions, the somatosensory evoked potential (SEP) was examined in

anaesthetised rats. When applied for 5 min EDA (25mM) produced an initial short-

lasting increase in the amplitude of the negative (N) phase of the SEP (by 12 ± 7%, n

= 5), invariably followed by a decrease (by 30 ± 12%, n = 5) (Fig. 9A,B). A higher

concentration produced a much lower initial increase in the SEP but subsequently

decreased it by 74 ± 5% (n = 5) (Fig 9B). EDA at 5 mM did not affect the size or

shape of the SEPs (n = 5).

To examine the role of GABA-A receptors in these effects, bicuculline was also

applied topically at a concentration of 10µM for 15 min before and during the

17

application of EDA. This compound increased the overall amplitude of the SEP by

77 ± 34% (n = 7)(Fig. 9C), with the greatest change occurring in the positive (P)

phase of the potential. However, the application of EDA (in the continued presence

of bicuculline) still caused an immediate and complete abolition of the evoked

positive wave and a gradual decrease in the overall amplitude of the SEP by 22 ±

15% (n = 7). There was no significant difference in the degree of SEP depression

caused by EDA on its own or in the presence of bicuculline (two-tailed t test; P =

0.40).

4. DISCUSSION

Although several agents are available with which to explore the pharmacology

and physiological roles of GABA receptors, uptake and metabolism, there are few

with which to probe neuronal or behavioural responses to the selective release of

endogenous GABA. As noted in the Introduction EDA appeared to be an efficient

GABA-mimetic, partly by inducing GABA release and partly by direct activation of

GABAA receptors in the CNS (Perkins and Stone, 1980, 1982a; Blaxter and

Cottrell,1985; Bokisch et al., 1984; Lloyd et al., 1982a,b). The effects of EDA

extended to the modulation of benzodiazepine binding (Morgan and Stone, 1982;

Davies et al., 1982).

The ability of EDA to release GABA is a result of the uptake of EDA itself by a

nipecotic acid-sensitive transporter, possibly producing an exchange with GABA

since the GABA release is calcium-independent (Lloyd et al., 1982a,b). Here,

nipecotic acid increased the size and duration of the EDA effects on fepsps, implying

that the balance of nipecotic acid’s activity is to prevent re-uptake of GABA rather

18

than to reduce the initial uptake of EDA (since the EDA response would then be

reduced).

The observation that EDA could produce a long-lasting depression in some slices

could imply a modification of receptors – such as those for glutamate – involved in

the generation of long-term depression (LTD), or could simply reflect the

accumulation of EDA within the tissue. We consider the former explanation more

likely given that the early time course of recovery was normal, the plateau of

depression could be maintained for up to one hour, and that shorter (5min)

applications of EDA appeared more likely to produce this LTD. Present experiments

are being directed at further understanding of the mechanisms of the LTD.

Following the proposal that this GABA-mimetic activity might involve a complex

formed between EDA and bicarbonate (Stone and Perkins, 1984), strong supporting

evidence for this hypothesis was obtained by demonstrating, physiologically and

spectroscopically, the existence of EDA monocarbamate in solutions of EDA gassed

with 5%CO2 in oxygen (Curtis and Malik, 1984; Myers and Nelson, 1990). Ejecting

the active monocarbamate ion onto individual neurones resulted in a larger and more

rapid depression of firing than was seen with EDA alone.

However, there has been little detailed systematic or quantitative work on the

pharmacology of EDA and the present work was designed to pursue this. The profile

of extracellular and intracellular recordings is consistent with previous work showing

that somatic applications of GABA and EDA generate bicuculline-sensitive

hyperpolarisations, while dendritic applications produce bicuculline-sensitive

depolarisations (Blaxter and Cottrell, 1985). The intracellular depolarisation

presumably corresponds to the extracellularly recorded depression of fepsps.

19

Equally, the increase of fepsp size is most probably a result of the intracellularly

recorded hyperpolarisation, both effects being resistant to bicuculline.

These results, together with the reduced responses to EDA in lowered

extracellular bicarbonate therefore support the postulated activation of GABA

receptors by EDA, probably via an EDA-bicarbonate complex. Perhaps more

importantly, the present study has revealed an additional mechanism of action of

EDA. Structurally, EDA monocarbamate is similar not only to GABA, but also to β-

alanine (Fig. 10), and EDA can release preloaded β-alanine as well as GABA from

neurones (Davies et al., 1983). Indeed, EDA is a more effective inhibitor of β-alanine

uptake than of GABA uptake (Davies et al., 1982; Forster et al., 1981), possible

suggesting that EDA may be a better inhibitor of glial transport (β-alanine) than of

neuronal uptake (GABA) (Davies et al., 1982).

It is, therefore, intriguing to discover in this study that the effects of EDA are

closely mimicked by applications of β-alanine, an amino acid recognised primarily for

its ability to activate glycine receptors. Both EDA and β-alanine produced an initial

increase of fepsp size preceding the reduction in their size, and both also produced

intracellularly recorded depolarisations, often preceded by an underlying

hyperpolarisation, with no fade being seen in the presence of the agonist. In contrast

GABA itself was able to produce only a decrease of fepsp size and intracellular

depolarisation, with both the extracellular and intracellular responses exhibiting

marked fade after about several seconds of the GABA applications. However,

despite these similarities, the effects of EDA and β-alanine remain largely distinct

since the EDA-induced depression of fepsps and the intracellular depolarisation were

blocked to a large extent, though not completely, by bicuculline, whereas these

effects of β-alanine were blocked by strychnine.

20

The simplest interpretation of these data is that while EDA, almost certainly in the

form of its monocarbamate, is able to activate GABAA receptors, and β-alanine is

able to activate strychnine-sensitive, conventional, glycineA receptors, there is a third

site which is resistant to bicuculline and strychnine at which both EDA and β-alanine,

again in the form of their monocarbamates, produce hyperpolarisation when recorded

intracellularly, and the increased fepsp size when recorded extracellularly. This

response would correspond to the dendritic, bicuculline-resistant, hyperpolarisation

observed by Blaxter and Cottrell (1985) when GABA was applied locally to the

dendritic region. It has been proposed by others, based on a substantial amount of

evidence on uptake, release and actions, that β-alanine could function as an

independent neurotransmitter (DeFeudis and del Rio, 1977; Sandberg and Jacobson,

1981).

The activity of EDA at sites which are resistant to blockade by bicuculline was

noted both in hippocampal slices and at the somatosensory cortical surface in vivo.

In order to explain this bicuculline-resistant fraction of responses to EDA, two

hypotheses were considered and tested in the slice preparation. Firstly, it is

conceivable that EDA might form, in addition to the monocarbamate structure, a

dicarbamate analogue which would resemble adipic acid in structure (Fig. 10). Such

a complex might then activate classical excitatory amino acid receptors, or a distinct

dicarboxylic acid receptor. The former possibility was tested using a high

concentration of the general glutamate receptor antagonist kynurenic acid (Perkins

and Stone 1982b; Stone and Darlington 2002), but this compound failed to block the

depression of antidromically-induced potentials or the intracellularly recorded

depolarising responses to EDA, suggesting that activation of glutamate receptors

cannot account for the results.

21

A role for an adipic acid receptor was also tested by applying adipic acid itself to

the slices, where it did produce a depression of fepsps and an intracellular

depolarisation. Those effects were not blocked by either bicuculline or strychnine,

leaving open the possibility that EDA dicarbamate and adipic acid could act partly by

a novel receptor site or transporter. Neither the EDA receptor nor the adipic acid site

were blocked by α-aminoadipic acid, an antagonist at glutamate receptors (Biscoe et

al., 1977; Stone, 1979).

The blockade of EDA responses by DIDS strongly supports the involvement of

common ionic mechanisms between GABA and EDA. In addition, it may be that the

poor blockade of EDA effects by bicuculline reflects a direct action on GABA

receptors distal to the site of action of GABA itself, explaining the common

dependence on chloride but with recognisably distinct pharmacology. A direct action

of some kind is certainly needed to account for the GABA-like effects in the absence

of synaptic terminals (Morgan and Stone, 1982; Davies et al., 1982) and such an

action, rather than activation of sites sensitive to agents such as β-alanine or adipic

acid, could account for many of the present results.

The work in vivo provides an interesting comparison with the hippocampus in vitro.

The SEPs result from the depolarization of pyramidal cells by thalamo-cortical

afferents. A subsequent positive wave due to the spread of depolarization to

superficial cortical layers (I – III) is increased by the removal of tonic inhibition from

GABAergic interneurons in these superficial layers (Addae et al. 2007; Jellema et al,

2004; Reyes et al. 1998), confirming a role for endogenous GABA release. The

results provide strong support to the findings in vitro. The profile of the trendplot of

the SEPs shows an initial slight increase in the SEP size and a subsequent more

prolonged decrease followed by recovery. The profile is similar to that seen in some

22

of the in vitro recordings (e.g. Fig 3).The in vivo profile is consistent with the report

that depolarizing GABA responses are found in the superficial cortical layers where

topical EDA is likely to exert its initial effect (DeFazio et al., 2005). It is clear that EDA

is able to exert a profound inhibitory action on SEPs similar to the inhibition noted in

vitro. Also, and perhaps more importantly, the effects of EDA are at least partially

resistant to bicuculline, since they are not prevented by concentrations of bicuculline

that induce an increase of neuronal excitability sufficient to double the amplitude of

an SEP and to induce multiple PS and spontaneous epileptiform activity in the

neocortex. Indeed, this failure of bicuculline to block the EDA effects in vivo suggests

that the direct effects of EDA on GABA receptors or channels noted above is

relatively more important in vivo than it is in vitro.

Following the initial reports concerning EDA, it became clear that an interaction

with bicarbonate or carbon dioxide was required for the excitatory and toxic action of

β-N-methylamino-L-alanine (BMAA) at excitatory amino acid receptors (Myers and

Nelson, 1990). Also, high concentrations of neutral amino acids such as L-alanine

and L-proline, can activate glutamate receptors (Pace et al., 1992). The formation of

carbamates or similar complexes of these amino acids in bicarbonate-containing

media could generate molecules able to mimic glutamate, making them relevant to

excitotoxicity in physiological or pathological conditions in which their concentrations

are elevated.

23

REFERENCES

Addae, J.I., Evans, S.M., Ali, N., Stone, T.W., 2000. NMDA-induced changes in a

cortical network in vivo are prevented by AMPA. Brain Res. 869, 211-215

Addae, J. I., Ali, N., Youssef, F., Stone, T. W.,2007. AMPA receptor activation

reduces epileptiform activity in the rat neocortex. Brain Res. 1158, 151-157.

Alger, B.E., Nicoll, R.A., 1982. Pharmacological evidence for two kinds of GABA

receptor on rat hippocampal pyramidal cells studied in vitro. J. Physiol. (Lond.), 328,

125-141.

Barbier, A.J., Guenaneche, F., Lefebvre, R.A., 1989. Influence of GABA and

ethylenediamine in the guinea-pig duodenum. J. Autonom. Pharmacol. 9, 279-291.

Begg, M., Molleman, A., Parsons, M., 2002. Modulation of the release of endogenous

γ-aminobutyric acid by cannabinoids in the guinea-pig ileum. Europ. J. Pharmacol.

434, 87-94.

Biscoe, T.J., Evans, R.H., Francis, A.A., Martin, M.R., Watkins, J.C., 1977. D-alpha-

aminoadipate as a selective antagonist of amino acid-induced and synaptic excitation

of mammalian spinal neurons. Nature 270, 743-745.

Blaxter, T.J., Cottrell, G.A., 1985. Actions of GABA and ethylenediamine on CA1

pyramidal neurons of the rat hippocampus. Quart. J. Exp. Physiol. 70, 75-93.

24

Bokisch, A.J., Bold, J.M., Gardner, C.R., Perkins, M.N., Roberts, C.J., Stone, T.W.,

Walker, R.J., 1984. The action of gamma-aminobutyric acid, (GABA) and

ethylenediamine (EDA) on Limulus and Helix central neurons and rat cerebellar and

sympathetic-ganglion neurons. Gen. Pharmacol. 15, 497-504.

Bonnet, U., Bingmann, D., 1995. GABA-A-responses of CA3 neurons: contribution of

bicarbonate and of chloride extrusion mechanisms. Neuroreport 6, 700-704.

Chaudhuri, L., Ganguly, D.K., 1994. Involvement of gamma-aminobutyric-acid in the

stimulatory effect of metoclopramide on gastrointestinal motility. Life Sci. 54, 253-

259.

Curtis, D.R., Malik, R. 1984. Ethylenediamine – a GABA-mimetic. Trends Pharmacol.

Sci. 5, 458.

Davies, L.P., Hambley, J.W., Johnston, G.A.R., 1982. Ethylenediamine as a GABA

agonist - enhancement of diazepam binding and interaction with GABA receptors

and uptake sites. Neurosci. Lett. 29, 57-61.

Davies, L.P., Drew, C.A., Chow, S.C., Skerritt, J.H., Johnston, G.A.R., 1983.

relationships between ethylenediamine and GABA transport-systems in rat-brain

slices. Neurochem. Internat. 5, 57 –64.

25

DeFazio, R. A., Hablitz, J. J., 2005. Horizontal spread of activity in neocortical

inhibitory networks. Devel. Brain Res. 157, 83-92.

DeFeudis, F. V., del Rio, R.M., 1977. Is β-alanine an inhibitory neurotransmitter?

Gen. Pharmac. 8, 177–180.

Erdo, S.L., Kiss, B., Riesz, M., Szporny, L., 1986. Stimulus-evoked efflux of GABA

from preloaded slices of the rabbit oviduct. Europ. J. Pharmacol. 130, 295 –303.

Ferguson, A.L., Stone, T.W., 2008. Adenosine preconditioning against ouabain but

not against glutamate on evoked potentials in rat hippocampal slices. Europ. J.

Neurosci. 28, 2084-2098..

Forster, P., Lloyd, H.G.E., Morgan, P.F., Parker, M., Perkins, M. N., Stone TW.,

1981. Ethylenediamine acts upon GABA receptors and uptake sites. Brit. J.

Pharmacol. 74, 274p.

Grover, L.M., Lambert, N.A., Schwartzkroin, P.A., Teyler, T.J., 1993. Role of HCO3-

ions in depolarizing GABA(A) receptor-mediated responses in pyramidal cells of rat

hippocampus. J. Neurophysiol. 69, 1541-1555.

Jellema, T., Brunia, C. H., Wadman, W. J., 2004. Sequential activation of

microcircuits underlying somatosensory-evoked potentials in rat neocortex.

Neuroscience 129, 283-295.

26

Kerr, D.I.B., Ong, J., 1982. GABA actions in guinea-pig ileum through release of

endogenous GABA by ethylenediamine. Clin. Exp. Pharmacol. Physiol. 9, 390.

Kerr, D.I.B., Ong, J., 1984. Evidence that ethylenediamine acts in the isolated ileum

of the guinea-pig by releasing endogenous GABA. Brit. J. Pharmacol. 83, 169 –177.

Kerr, D.I.B., Ong, J., 1987. Bicarbonate-dependence of responses to

ethylenediamine in the guinea-pig isolated ileum - involvement of ethylenediamine-

monocarbamate. Brit. J. Pharmacol. 90, 763-769.

Lindquist, C.E.L., Laver, D.R., Birnir, B., 2005. The mechanism of SR95531 inhibition

at GABA(A) receptors examined in human α1β1 and α1β1γ2S receptors. J.

Neurochem. 94, 491-501.

Lloyd, H.G.E., Perkins, M.N., Gaitonde, M.K., Stone, T.W., 1982a. Uptake and

calcium-dependent release of ethylenediamine (1,2-diaminoethane) by rat-brain

slices. J. Neurochem. 38, 1118-1122

Lloyd, H.G.E., Perkins, M.N., Stone, T.W., 1982b. Ethylenediamine as a specific

releasing agent of gamma-aminobutyric acid in rat striatal slices. J. Neurochem. 38,

1168-1169.

Maggi, C.A., Giuliani, S., Santicioli, P., Selleri, S., Meli, A., 1989. The motor response

to ethylenediamine of the rat isolated duodenum - involvement of GABAergic

transmission. Naunyn-Schmied. Archiv.Pharmacol. 340, 419 –423.

27

Mckay, A., Krantis, A., 1991. The effects of ethylenediamine in the rat small-intestine

- a powerful relaxant of the muscularis. Canad. J. Physiol. Pharmacol. 69, 199-204.

Morgan, P.F., Stone, T.W., 1982. Ethylenediamine and GABA potentiation of

[3H]diazepam binding to benzodiazepine receptors in rat cerebral cortex. J.

Neurochem. 39, 1446-1451.

Myers, T.G., Nelson, S.D., 1990. Neuroactive carbamate adducts of β-N-

methylamino-l-alanine and ethylenediamine - detection and quantitation under

physiological conditions by C-13 NMR. J. Biol. Chem. 265, 10193-10195.

Ong, J., Kerr, D.I.B., 1987. Comparison of GABA-induced responses in various

segments of the guinea-pig intestine. Europ. J. Pharmacol. 134, 349-353.

Pace, J.R., Martin, B.M., Paul, S.M., Rogawski, M.A., 1992. High concentrations of

neutral amino acids activate NMDA receptor currents in rat hippocampal neurons.

Neurosci. Lett. 141, 97-100.

Perkins, M.N., Stone, T.W., 1980. Aminophylline and theophylline derivatives as

antagonists of neuronal depression by adenosine - a microiontophoretic study.

Archiv. Internat. Pharmacodyn. Therapie 246, 205 –214.

28

Perkins, M.N., Stone, T.W., 1982a. Comparison of the effects of ethylenediamine

analogs and gamma-aminobutyric acid on cortical and pallidal neurons. Brit. J.

Pharmacol. 75, 93-99.

Perkins MN, Stone TW , 1982b. An iontophoretic investigation of the action of

convulsant kynurenines and their interaction with the endogenous excitant quinolinic

acid. Brain Res. 247, 184-187.

Perkins, M.N., Bowery, N.G., Hill, D.R., Stone, T.W., 1981. Neuronal responses to

ethylenediamine - preferential blockade by bicuculline. Neurosci. Lett. 23, 325-327.

Reyes, A., Lujan, R., Rozov, A., Burnashev, N., Somogyi, P., Sakmann, B., 1998.

Target-cell-specific facilitation and depression in neocortical circuits. Nature

Neurosci. 1, 279-285.

Sandberg, M., Jacobson, I., 1981. β-alanine, a possible neurotransmitter in the visual

system. J. Neurochem. 37, 1353-1356.

Sarthy, P.V., 1983. Release of [H-3] gamma-aminobutyric acid from glial (Muller)

cells of the rat retina - effects of K+, veratridine, and ethylenediamine. J. Neurosci. 3,

2494-2503.

Stone, T. W., 1979. Selective antagonism of amino acids by α-aminoadipate on

pyramidal tract neurons but not Purkinje cells. Brain Res., 166, 217-220.

29

Stone, T.W., 2001. Kynurenines in the CNS - from obscurity to therapeutic

importance. Progr. Neurobiol. 64, 185-218.

Stone, T.W., 2007. Kynurenic acid blocks nicotinic synaptic transmission to

hippocampal interneurons in young rats. Europ. J. Neurosci. 25, 2656 – 2665

Stone, T.W., Darlington, L.G., 2002. Endogenous kynurenines as targets for drug

discovery and development. Nature Rev. Drug Disc. 1, 609-620

Stone, T.W., Perkins, M.N., 1979. Is adenosine the mediator of opiate actions on

neuronal firing rate? Nature 281, 227-228.

Stone, T.W., Perkins, M.N., 1984. Ethylenediamine as a GABA-mimetic. Trends

Pharmacol. Sci. 5, 241-243.

30

Figure legends:

Figure 1. Responses of hippocampal slices to EDA.

A. illustrates the effects of EDA when applied for 10 min to allow the attainment of a

plateau depression. Responses to three concentrations are illustrated: 100µM (n =

5), 300µM (n = 4) and 1000µM (n = 16). B. is of field epsps from one slice showing

the baseline potential (the largest epsp) and sample potentials recorded at the peak

depression induced by 50, 100, 300, 500 and 1000µM EDA (progressively smaller

epsps). The traces indicate that the depression of the epsps was not accompanied

by any change of axonal volley. C. shows the log concentration-response curves for

EDA (filled circles, n = 6-8) and GABA (open circles, n = 5-6). The measurements

were based on the mean of the last 10 epsps recorded during superfusion with EDA

or GABA for 10 min. D. shows the mean data for 4 slices in which EDA (1mM)

produced a long-lasting depression of the fepsp.

Calibrations for B: 0.5mV and 5ms.

Data are shown as mean ± S.E.M..

Figure 2. Effects of bicuculline and gabazine on responses to EDA.

A. The time course of the depression of fepsps by EDA (1000µM, n = 5) applied for 2

min, and its partial blockade by bicuculline methobromide (BIC, 20µM, n = 4).

B. illustrates records of epsps from a slice treated with bicuculline 100µM, showing

the baseline potential, the depression produced by EDA 1000µM, and the partial

blockade by bicuculline.

C. shows the overall data in which bicuculline at 20µM (n = 5) and 100µM (n = 3) or

gabazine at 20µM (n = 3) and 100µM (n = 3) produced a partial blockade of the

31

responses to EDA (1mM). There is no significant difference between the blockade

produced by the two concentrations of antagonist, suggesting that the blockade is

maximal. Strychnine at 10 or 100µM failed to modify the responses to EDA.

Data are shown as mean ± S.E.M..

** P < 0.01 compared with the control response to EDA (Dunnett’s test for multiple

comparisons).

Figure 3. Effects of nipecotic acid on responses to EDA.

A. Amplitudes of the first of a series of fepsp depressions induced by EDA 300µM

(filled circles), the increase in amplitude and duration of the response sampled as the

maximum response to EDA in the presence of nipecotic acid 100µM (open circles),

and recovery towards the original response profile (filled triangles). Nipecotic acid

was applied for 15 min before, during and for 30 min following EDA.

B. Sample traces showing the fepsps recorded (a) at baseline, (b) at the peak of

depression produced by the first application of EDA and (c) at the peak of depression

produced by EDA in the presence of nipecotic acid.

C. summarises the time course of change in the mean size of the depression

produced by EDA when slices were superfused with nipecotic acid (100µM).

D. summarises the overall data in which nipecotic acid produced a significant

increase in the amplitude of the EDA response.

Data are shown as mean ± S.E.M. (n = 6).

** P < 0.01 compared with the control response to EDA.

32

Figure 4. Effects of kynurenic acid on antidromically-evoked spikes.

A. The depression of antidromic spike amplitude produced by EDA 1mM in the

absence (filled circles) and presence (open circles) of kynurenic acid 2mM.

Kynurenic acid was present for 15 min before, during and for 20 min following the

application of EDA, but failed to modify the EDA response.

B. Sample antidromic spikes recorded (a) at baseline, (b) at the peak depression

produced by EDA and (c) at the peak depression produced by EDA in the presence

of kynurenic acid.

Data are shown as mean ± S.E.M. (n = 5).

Figure 5. Effects of β-alanine on fepsps.

A. When applied for 5min at concentrations of 500µM, β-alanine (closed circles, n =

4) induced a depression of fepsp slope which was smaller and slower than that

produced by EDA. Strychnine 100µM (open triangles, n = 4), applied for 15 min

before, during and for 30 min after β-alanine reduced the size of depressant

response but did not change the initial increase in fepsp size.

B. illustrates sample fepsps at baseline (a), at the time of maximum inhibition by β-

alanine (b) and when blocked by strychnine (c).

C. chart summarising the decreased fepsp size produced by β-alanine (β-ala), the

failure of bicuculline (bic) to modify that depression, and the blockade by strychnine

(stry), and failure of bicuculline or strychnine to modify the initial increase in fepsp

size produced by β-alanine.

** P < 0.01 (n = 4) compared with the control response to β-alanine alone.

Calibrations: 1mV, 10ms.

33

Figure 6. Intracellular recordings of the effects of EDA on a CA1 pyramidal neuron.

A. illustrates a typical record of the depolarisation produced by EDA 300µM. The

depolarisation is associated with an increase in membrane conductance, tested

using 500ms pulses of hyperpolarising current (downward deflections). The upward

deflections on the trace are spontaneous epsps which are suppressed by the

membrane depolarisation. B. illustrates both the larger depolarisation and

conductance change produced at 1mM EDA, and an underlying hyperpolarisation

which occurs before and following the depolarisation (6 neurons). C. illustrates a

neuron depolarised by EDA (500µM) in which, during the period indicated by the

hatched bar, the membrane potential was manually clamped at the resting potential

to confirm that the decrease of membrane resistance was still observed under these

conditions.

Calibrations in A: 5mV and 1min; calibrations in B, C: 10mV and 1 min.

Figure 7. Blockade of EDA depolarisation by bicuculline.

A. (a) illustrates the initial depolarisation induced by EDA, 1mM, while (b) shows the

response after the application of bicuculline methobromide (20µM, 20min). In (c) the

response to EDA recovers substantially following removal of the antagonist. B. The

effect of bicuculline was statistically significant, with recovery reaching an amplitude

not significantly different from the initial value (n = 4). Calibrations in A: 20mV and

2min. Action potentials triggered in this cell have been cut off.

**P < 0.01 relative to EDA alone.

34

Figure 8. Responses to EDA and adipic acid.

A. Control responses are shown (open circles) for typical responses to an application

of EDA at 500µM followed by the smaller and slower effect of adipic acid (2mM).

Superimposed on that plot are responses after superfusion with α-aminoadipic acid

(1mM, closed circles). There is no effect of α-aminoadipic acid on the response to

adipic acid, although there is a small apparent reduction in the response to EDA.

B. The overall analysis indicated that α-aminoadipic acid had no significant effect on

the response to EDA or adipic acid. Data are shown as mean ± S.E.M. for n = 4.

Figure 9. Effects of EDA on somatosensory evoked potentials (SEPs) in vivo.

A. A typical individual record showing a response to topical EDA (25 mM for 5 min)

causing an initial increase and then a decrease in the SEPs. B. Graphs of the pooled

data showing a concentration dependent effect of EDA at 25 mM (black circles) and

50 mM (open circles) (n = 5 at each concentration). The initial increase (more

marked at the lower concentration) was followed by the decrease which recovered

slowly to the control level. C. Bicuculline (10 µM for 15 min) caused an increase in

the SEPs but did not significantly affect the decrease in the size of the SEP caused

by EDA. Note especially the immediate and complete abolition by EDA of the

increased positive wave caused by bicuculline. The graphs show the changes in

amplitude of the first negative (downward) wave of the SEP with each point being an

average of 30 SEPs recorded over 1 min. The representative waveforms represent

individual sample SEPs taken at the times (a,b,c,d) indicated on the time graph.

Figure 10 Chemical structures of relevant molecules.

35

Time (min)0 10 20 30 4

feps

p sl

ope

(% b

asel

ine)

00

20

40

60

80

100

B

1mM basal

1

12A 0

1000µM300µM100µM

EDA

log concentration (µM)101 102 103 104

inhi

bitio

n of

feps

p (%

)

0

20

40

60

80

100

EDA

GABA

T i m e (m i n)0 10 20 30

feps

p sl

ope

(% b

asel

ine)

0

20

40

60

80

100

120E D A

DC

5ms

0.5mV

Figure 1

40

36

Figure 2

A

Time (min)0 5 10 15 20 25 30 35

feps

p sl

ope

(% b

asel

ine)

0

20

40

60

80

100

120

control+BIC 20µMrecovery

EDA

B

10ms

1mV

C

EDA + B I C +gabazine strychninecon 20 100 20 100 10 100 µM

ED

A in

hibi

tion

of fE

PS

P (%

)

0

20

40

60

80

100

** **

37

3

Time min)0 5 10 15 20 25 30 35

feps

p sl

ope

(% b

asel

ine)

0

20

40

60

80

100

120

14A 0

control+nipecotic acid recovery

EDA

feps

p de

pres

sion

of f

epsp

(%)

0

20

40

60

80

100B

10ms

1mV

c b a

EDA Econtrol nipe

inhi

bitio

n of

feps

p sl

ope

(%)

0

20

40

60

80

100

*D

(

C

1 2 3 4 5 6 7 8 9 EDA response number

nipecotic acid

*

Figure

DA+ EDAcotic recovery

38

Figure 4

Time (min)0 5 10 15 20 25an

tidro

mic

spi

ke a

mpl

itude

(% b

asel

ine)

0

20

40

60

80

100

120

EDA controlEDA+ kynurenic acid

EDA

A B

10ms

1mV

c b a

39

Figure 5

160 β-alanine B

A

Time (min)0 10 20 30 40 5

feps

p sl

ope

(% b

asel

ine)

00

20

40

60

80

100

120

140

β-alanineβ-alanine + strychnine

sp (%

)

100

C

decrease of fepsp increase of fepspβ-ala β-ala β-ala β-ala β-ala β-ala + bic + stry + bic + stry

chan

ge fr

om b

asel

ine

fep

0

20

40

60

80

**

FIGURE 5

1mV

10ms

b c a

40

Figure 6

41

Figure 7

-6.0

-6.5

-7.0

-7.5

-8.0

-8.5

-9.0

-9.5

-10.0

-10.5

-11.0

-11.5

-12.0

-12.5

-13.0

-13.5

mV

ADC 0

200

2min20mV

EDA

A B

EDA EDA+ EDA

depo

laris

atio

n (m

V)

0

5

10

15

20

25

30

**

control bicuculline recovery

a b c

42

Figure 8

time (min)0 20 40 60 80 100 1

feps

p sl

ope

(% b

asel

ine)

200

20

40

60

80

100

12A

0

with α−aminoadipate controls

EDA AA

EDA EDA+ AdA AdA+ αAA αAA

depr

essi

on o

f fep

sp s

lope

0

20

40

60

80

100

B

43

Figure 9

Time (min)0 10 20 30 40 50 6

SEP

am

plitu

de (m

V)

00

1

2

3

4 2mV

20ms

cb

ab

d

c

a

A

Time (min)0 20 40 60 80 10

SE

P a

mpl

itude

(%)

00

20406080

100120140

EDA 25mM EDA 50mM

EDA

B

Time (min)0 20 40 60 80 100 120S

EP

am

plitu

de (%

)

0

50

100

150

200

250bic

EDA 20ms

1mV

a dcb

d

cb

a

C

44

Figure 10

H2N

OH

O

OH

O

H2NNH2

H2N

OH

O

H2NNH2+

H2NNH

-OOH

OOH

O

HO

O

HN

NH

OH

O

HO

O

4-aminobutyric acid (GABA) β−alanine

ethylenediamine

ethylenediamine bicarbonate complex ethylenediamine carbamate

adipic acid ethylenediamine di-carbamate

45